Embodiment in Human and AI-generated Data Propositions

Many readers struggle to interpret the truth value of claims that utilize numerical data in mathematical expressions. These struggles can arise from a lack of context for the data, which can lead to misinterpretation of meaning or overattribution of truth, neither of which serves readers in understanding data-driven concepts like “food insecurity.” Some readers do not experience concepts like food insecurity directly, and hence, those concepts become real (or not) through the data used to express them. This research shows how writers assert the meaning of data about food insecurity by grounding it in an embodied sense of what that data means. This writerly approach is contrasted with examples of artificial intelligence-generated claims about “food insecurity” that are less commonly grounded in an embodied sense of truth. The findings are used to suggest what human writers uniquely add to data communication: an embodied, experiential sense of what the data mean.